Research on Optimization and Tool Wear Prediction of Ultrasonic Vibration-Assisted Blade Dicing System = 초음파 진동 보조 블레이드 다이싱 시스템의 최적화 및 공구 마모 예측 연구

다국어 초록 (Multilingual Abstract)

With the rapid development of the semiconductor, solar cells, and micro electro mechanical systems （MEMS） industries, the demand for high-quality dicing of single crystal silicon wafers is growing. Ultrasonic vibration-assisted blade dicing （UVABD） technology has become a hot topic in the field of precision machining due to its ability to significantly reduce cutting forces, improve surface integrity, and extend tool life. However, the performance stability of UVABD systems is limited by the bonding process of ultrasonic transducers, and the complex mechanism of tool wear during machining leads to fluctuations in quality, and difficulties in predicting tool life. This thesis focuses on the key technologies of UVABD systems, including two core components: first, the optimization of the fundamental performance of the ultrasonic vibration system; second, the investigation of tool wear mechanisms and the development of a tool life prediction model during the dicing process. First, addressing the issue of unstable bonding performance between the Lead zirconate titanate （PZT） and the substrate in the ultrasonic transducer, this study compared the performance of three adhesives: Cement glue, Super glue, and Hysol. Through impedance analysis and vibration amplitude test, Hysol was determined to be the optimal bonding solution, as its amplitude at 120 V was higher than that of Super glue and Cement glue, providing a stable vibration source for subsequent processing. Second, a UVABD tool wear model was established that accounts for grain morphology and tool motion. Based on tool surface topography measured by a white light interferometer, a mathematical model for grain protrusion height was developed; by analyzing the dynamic grain motion under ultrasonic vibration, a wear volume formula related to the normal force on the grain and the effective sliding distance was established; simultaneously, considering the influence of radial vibration on chip formation, the maximum undeformed chip thickness was derived, thereby obtaining the average grain wear evolution trend and tool life prediction model at a given critical wear height. Experimental validation indicates that, within the 0–120-minute range, the relative error between model predictions and experimental results is 9%–11%, accurately reflecting the evolution from the initial wear stage to the stable wear stage. Furthermore, the study found that wafer edge integrity decreases with tool wear, and high feed rates and large dicing depths significantly increase chipping dimensions. These findings reveal the relationship among UVABD kinematics, tool wear evolution, and tool life, providing a crucial theoretical and engineering basis for parameter optimization, tool life prediction, and machining quality control in the machining of hard and brittle materials. Keywords: Ultrasonic machining; Tool wear; Life prediction; Modeling; PZT bonding; Single-crystal silicon wafers

목차 (Table of Contents)

• Chapter 1 Introduction 1
• 1.1 Background and significance 1
• 1.1.1 Background and challenges of precision machining of hard and brittle materials 1
• 1.1.2 Development and challenges of UVAM technology 3
• 1.1.3 Tool wear modeling and life prediction 4
• 1.1.4 Significance of this study 6
• 1.2 Literature review 7
• 1.2.1 Research on ultrasonic transducer bonding 7
• 1.2.2 Research on the mechanism of ultrasonic-assisted machining 8
• 1.2.3 Research on tool wear modeling and life prediction 9
• 1.2.4 Limitations of existing research and the focus of this study 11
• 1.3 Research content and technical approach 12
• 1.3.1 Research content 12
• 1.3.2 Technical approach 13
• 1.4 Thesis structure 14
• 1.5 Discussion 16
• Chapter 2 Optimization of adhesive performance 18
• 2.1 Introduction 18
• 2.2 Mechanics of the bonding interface 19
• 2.2.1 PZT coupling and stress transfer mechanisms 19
• 2.2.2 Damping characteristics of the adhesives 20
• 2.2.3 Resonance frequency and system stiffness 21
• 2.3 Experimental materials 22
• 2.3.1 Selection of PZT 22
• 2.3.2 Substrate material 23
• 2.3.3 Selection of adhesives 25
• 2.4 Experimental methods 26
• 2.4.1 Hardness test 26
• 2.4.2 Impedance test 28
• 2.4.3 Amplitude test 30
• 2.4.4 Experimental error control 32
• 2.5 Results 32
• 2.5.1 Hardness and analysis of adhesives 32
• 2.5.2 Impedance analysis 34
• 2.5.3 Vibration amplitude analysis 36
• 2.5.4 Impact on UVABD machining stability 43
• 2.6 Discussion 43
• Chapter 3 Kinematics of UVABD and grain interaction models 45
• 3.1 Introduction 45
• 3.2 Model assumptions and modeling of tool surface topography 47
• 3.2.1 Modeling assumptions 47
• 3.2.2 Modeling grain size and protrusion height distributions 48
• 3.2.3 Determination of active grains 52
• 3.3 Kinematic analysis under ultrasonic vibration 54
• 3.4 Model for maximum undeformed chip thickness 57
• 3.5 Discussion 60
• Chapter 4 Mechanisms of tool wear and tool life prediction model 63
• 4.1 Introduction 63
• 4.2 Volume model of single-grain wear 65
• 4.2.1 Assumptions 65
• 4.2.2 Calculation of normal force and sliding distance 66
• 4.2.3 Formula for the evolution of wear height 68
• 4.3 Tool life prediction 69
• 4.4 Discussion 70
• Chapter 5 Experimental validation and analysis of surface integrity 73
• 5.1 Introduction 73
• 5.2 Experimental design and conditions 74
• 5.2.1 Workpiece and tool parameters 74
• 5.2.2 Experimental setup and procedure 76
• 5.2.3 Measurement and data processing 77
• 5.3 Validation of the tool wear model 78
• 5.3.1 Wear morphology at different times 78
• 5.3.2 Comparison of wear prediction with experimental results 81
• 5.3.3 Influence of machining parameters on wear 83
• 5.4 Analysis of surface integrity and chipping 87
• 5.4.1 The effect of tool wear on edge chipping 87
• 5.4.2 Effect of dicing parameters on surface integrity 90
• 5.5 Discussion 94
• Chapter 6 Conclusions and outlook 96
• 6.1 Main conclusions 96
• 6.2 Key innovations 99
• 6.3 Limitations 101
• 6.4 Future research directions 103
• Reference 106
• Appendices 115
• Abstracts 116